US20100046738A1  Method and apparatus for DMT crosstalk cancellation  Google Patents
Method and apparatus for DMT crosstalk cancellation Download PDFInfo
 Publication number
 US20100046738A1 US20100046738A1 US12/583,694 US58369409A US2010046738A1 US 20100046738 A1 US20100046738 A1 US 20100046738A1 US 58369409 A US58369409 A US 58369409A US 2010046738 A1 US2010046738 A1 US 2010046738A1
 Authority
 US
 United States
 Prior art keywords
 symbol
 channel
 pseudo
 victim
 khz
 Prior art date
 Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
 Granted
Links
 238000004891 communication Methods 0 abstract claims description 24
 230000001747 exhibited Effects 0 abstract claims description 19
 230000000875 corresponding Effects 0 abstract claims description 15
 230000000051 modifying Effects 0 abstract claims description 15
 230000002829 reduced Effects 0 abstract claims description 9
 125000002103 4,4'dimethoxytriphenylmethyl group Chemical group data:image/svg+xml;base64,<?xml version='1.0' encoding='iso-8859-1'?>
<svg version='1.1' baseProfile='full'
              xmlns='http://www.w3.org/2000/svg'
                      xmlns:rdkit='http://www.rdkit.org/xml'
                      xmlns:xlink='http://www.w3.org/1999/xlink'
                  xml:space='preserve'
width='300px' height='300px' >
<!-- END OF HEADER -->
<rect style='opacity:1.0;fill:#FFFFFF;stroke:none' width='300' height='300' x='0' y='0'> </rect>
<path class='bond-0' d='M 157.005,210.626 129.983,197.722' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-0' d='M 155.533,203.286 136.618,194.253' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-23' d='M 157.005,210.626 181.692,193.677' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-1' d='M 129.983,197.722 127.648,167.868' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-2' d='M 127.648,167.868 152.335,150.919' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-2' d='M 134.741,170.263 152.022,158.399' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-3' d='M 152.335,150.919 179.357,163.824' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-5' d='M 152.335,150.919 150,121.065' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-4' d='M 179.357,163.824 181.692,193.677' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-4' d='M 173.736,168.769 175.371,189.666' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-6' d='M 150,121.065 149.028,108.634' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-6' d='M 149.028,108.634 148.055,96.2024' style='fill:none;fill-rule:evenodd;stroke:#7F7F7F;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-7' d='M 150,121.065 120.146,123.4' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-15' d='M 150,121.065 179.854,118.73' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-8' d='M 120.146,123.4 103.197,98.7138' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-8' d='M 112.667,123.087 100.802,105.807' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-24' d='M 120.146,123.4 107.242,150.422' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-9' d='M 103.197,98.7138 73.3437,101.049' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-10' d='M 73.3437,101.049 60.439,128.07' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-10' d='M 76.8123,107.683 67.779,126.598' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-11' d='M 60.439,128.07 47.8428,129.056' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-11' d='M 47.8428,129.056 35.2466,130.041' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-13' d='M 60.439,128.07 77.388,152.757' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-12' d='M 27.1588,125.414 20.3976,115.567' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-12' d='M 20.3976,115.567 13.6364,105.719' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-14' d='M 77.388,152.757 107.242,150.422' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-14' d='M 81.3991,146.436 102.297,144.801' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-16' d='M 179.854,118.73 196.803,143.417' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-16' d='M 187.333,119.043 199.198,136.324' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-25' d='M 179.854,118.73 192.758,91.7087' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-17' d='M 196.803,143.417 226.656,141.082' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-18' d='M 226.656,141.082 239.561,114.06' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-18' d='M 223.188,134.448 232.221,115.533' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-19' d='M 239.561,114.06 252.157,113.075' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-19' d='M 252.157,113.075 264.753,112.09' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-21' d='M 239.561,114.06 222.612,89.3737' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-20' d='M 272.841,116.716 279.602,126.564' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-20' d='M 279.602,126.564 286.364,136.412' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-22' d='M 222.612,89.3737 192.758,91.7087' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-22' d='M 218.601,95.6947 197.703,97.3292' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<text x='145.334' y='96.2024' style='font-size:9px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#7F7F7F' ><tspan>*</tspan></text>
<text x='25.9241' y='135.396' style='font-size:9px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#FF0000' ><tspan>O</tspan></text>
<text x='264.753' y='116.716' style='font-size:9px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#FF0000' ><tspan>O</tspan></text>
</svg>
 data:image/svg+xml;base64,<?xml version='1.0' encoding='iso-8859-1'?>
<svg version='1.1' baseProfile='full'
              xmlns='http://www.w3.org/2000/svg'
                      xmlns:rdkit='http://www.rdkit.org/xml'
                      xmlns:xlink='http://www.w3.org/1999/xlink'
                  xml:space='preserve'
width='85px' height='85px' >
<!-- END OF HEADER -->
<rect style='opacity:1.0;fill:#FFFFFF;stroke:none' width='85' height='85' x='0' y='0'> </rect>
<path class='bond-0' d='M 43.9848,59.1774 36.3287,55.5211' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-0' d='M 43.5676,57.0978 38.2083,54.5383' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-23' d='M 43.9848,59.1774 50.9793,54.3752' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-1' d='M 36.3287,55.5211 35.6671,47.0626' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-2' d='M 35.6671,47.0626 42.6616,42.2604' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-2' d='M 37.6767,47.7412 42.5729,44.3796' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-3' d='M 42.6616,42.2604 50.3177,45.9167' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-5' d='M 42.6616,42.2604 42,33.8018' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-4' d='M 50.3177,45.9167 50.9793,54.3752' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-4' d='M 48.7252,47.3178 49.1883,53.2388' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-6' d='M 42,33.8018 41.7245,30.2796' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-6' d='M 41.7245,30.2796 41.449,26.7574' style='fill:none;fill-rule:evenodd;stroke:#7F7F7F;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-7' d='M 42,33.8018 33.5415,34.4634' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-15' d='M 42,33.8018 50.4585,33.1402' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-8' d='M 33.5415,34.4634 28.7392,27.4689' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-8' d='M 31.4222,34.3747 28.0607,29.4785' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-24' d='M 33.5415,34.4634 29.8851,42.1195' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-9' d='M 28.7392,27.4689 20.2807,28.1305' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-10' d='M 20.2807,28.1305 16.6244,35.7866' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-10' d='M 21.2635,30.0102 18.7041,35.3694' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-11' d='M 16.6244,35.7866 13.0555,36.0657' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-11' d='M 13.0555,36.0657 9.48655,36.3449' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-13' d='M 16.6244,35.7866 21.4266,42.7811' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-12' d='M 7.195,35.0341 5.27932,32.2439' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-12' d='M 5.27932,32.2439 3.36364,29.4537' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-14' d='M 21.4266,42.7811 29.8851,42.1195' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-14' d='M 22.5631,40.9902 28.484,40.527' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-16' d='M 50.4585,33.1402 55.2608,40.1348' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-16' d='M 52.5778,33.229 55.9393,38.1251' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-25' d='M 50.4585,33.1402 54.1149,25.4841' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-17' d='M 55.2608,40.1348 63.7193,39.4732' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-18' d='M 63.7193,39.4732 67.3756,31.8171' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-18' d='M 62.7365,37.5935 65.2959,32.2342' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-19' d='M 67.3756,31.8171 70.9445,31.5379' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-19' d='M 70.9445,31.5379 74.5134,31.2588' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-21' d='M 67.3756,31.8171 62.5734,24.8226' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-20' d='M 76.805,32.5696 78.7207,35.3598' style='fill:none;fill-rule:evenodd;stroke:#FF0000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-20' d='M 78.7207,35.3598 80.6364,38.15' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-22' d='M 62.5734,24.8226 54.1149,25.4841' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<path class='bond-22' d='M 61.4369,26.6135 55.516,27.0766' style='fill:none;fill-rule:evenodd;stroke:#000000;stroke-width:2px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity:1' />
<text x='40.6781' y='26.7574' style='font-size:2px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#7F7F7F' ><tspan>*</tspan></text>
<text x='6.84516' y='37.8622' style='font-size:2px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#FF0000' ><tspan>O</tspan></text>
<text x='74.5134' y='32.5696' style='font-size:2px;font-style:normal;font-weight:normal;fill-opacity:1;stroke:none;font-family:sans-serif;text-anchor:start;fill:#FF0000' ><tspan>O</tspan></text>
</svg>
 [H]C1=C([H])C([H])=C(C([H])=C1[H])C(*)(C1=C([H])C([H])=C(OC([H])([H])[H])C([H])=C1[H])C1=C([H])C([H])=C(OC([H])([H])[H])C([H])=C1[H] 0 title 1
 230000001131 transforming Effects 0 claims 8
 206010063834 Oversensing Diseases 0 description 108
 238000000034 methods Methods 0 description 42
 239000011230 binding agents Substances 0 description 8
 238000002592 echocardiography Methods 0 description 8
 230000001808 coupling Effects 0 description 6
 238000010168 coupling process Methods 0 description 6
 238000005859 coupling reaction Methods 0 description 6
 230000004044 response Effects 0 description 6
 238000004088 simulation Methods 0 description 6
 230000001702 transmitter Effects 0 description 5
 230000001360 synchronised Effects 0 description 4
 102100015879 ADSL Human genes 0 description 3
 101700073394 PUR8 family Proteins 0 description 3
 238000004364 calculation methods Methods 0 description 3
 230000000694 effects Effects 0 description 3
 230000014509 gene expression Effects 0 description 3
 230000000670 limiting Effects 0 description 3
 239000002609 media Substances 0 description 3
 102100003568 SDSL Human genes 0 description 2
 101700082902 SDSL family Proteins 0 description 2
 1 VDSL Proteins 0 description 2
 238000006011 modification Methods 0 description 2
 230000004048 modification Effects 0 description 2
 230000001603 reducing Effects 0 description 2
 238000001228 spectrum Methods 0 description 2
 241000659073 Achaea echo Species 0 description 1
 101700042928 LIN2 family Proteins 0 description 1
 238000004220 aggregation Methods 0 description 1
 230000002776 aggregation Effects 0 description 1
 238000004458 analytical methods Methods 0 description 1
 230000001721 combination Effects 0 description 1
 230000001934 delay Effects 0 description 1
 238000005516 engineering processes Methods 0 description 1
 239000000835 fiber Substances 0 description 1
 230000001976 improved Effects 0 description 1
 230000001939 inductive effects Effects 0 description 1
 238000009434 installation Methods 0 description 1
 239000011159 matrix materials Substances 0 description 1
 230000036629 mind Effects 0 description 1
 238000006722 reduction reaction Methods 0 description 1
 238000000926 separation method Methods 0 description 1
Classifications

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04B—TRANSMISSION
 H04B3/00—Line transmission systems
 H04B3/02—Details
 H04B3/32—Reducing crosstalk, e.g. by compensating

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L25/00—Baseband systems
 H04L25/02—Details ; Arrangements for supplying electrical power along data transmission lines
 H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks ; Receiver end arrangements for processing baseband signals
 H04L25/03006—Arrangements for removing intersymbol interference
 H04L25/03343—Arrangements at the transmitter end

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04M—TELEPHONIC COMMUNICATION
 H04M11/00—Telephonic communication systems adapted for combination with other electrical systems
 H04M11/06—Simultaneous speech and telegraphic or other data transmission over the same conductors
 H04M11/062—Simultaneous speech and telegraphic or other data transmission over the same conductors using different frequency bands for speech and other data

 H—ELECTRICITY
 H04—ELECTRIC COMMUNICATION TECHNIQUE
 H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
 H04L25/00—Baseband systems
 H04L25/02—Details ; Arrangements for supplying electrical power along data transmission lines
 H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks ; Receiver end arrangements for processing baseband signals
 H04L25/03006—Arrangements for removing intersymbol interference
 H04L2025/0335—Arrangements for removing intersymbol interference characterised by the type of transmission
 H04L2025/03426—Arrangements for removing intersymbol interference characterised by the type of transmission transmission using multipleinput and multipleoutput channels
Abstract
Description
 This application claims the benefit of prior filed copending Provisional Applications: No. 61/091,358 filed on Aug. 23, 2008 entitled “Frequency Domain Crosstalk Canceling with Different Symbol Rates” (Attorney Docket: VELCP086P) and No. 61/151,441 filed on Feb. 10, 2009 entitled “Frequency Domain Crosstalk Canceling with Different Symbol Rates” (Attorney Docket: VELCP088P) which are incorporated herein by reference in their entirety as if fully set forth herein.
 1. Field of Invention
 The field of the present invention relates to multitone transceivers.
 2. Description of the Related Art
 Digital Subscriber Lines (DSL) technology and improvements thereon including: G.Lite, ADSL, VDSL, HDSL all of which are broadly identified as XDSL have been developed to increase the effective bandwidth of existing subscriber line connections, without requiring the installation of new fiber optic cable. An XDSL modem operates at frequencies higher than the voice band frequencies, thus an XDSL modem may operate simultaneously with a voice band modem or a telephone conversation. Currently there are over ten discrete XDSL standards, including: G.Lite, ADSL, VDSL, SDSL, MDSL, RADSL, HDSL, etc. Each standard is typically implemented with a multitone (DMT) line code, or modulation protocol.
 The primary factor limiting the bandwidth or channel capacity of any of the above discussed XDSL protocols is noise, whether that noise be from echo, channel cross talk, impulse or background sources. Efforts are made throughout the DSL architecture to minimize noise.
 For voice applications a certain amount of echo was considered a positive feedback for telephone conversations until the longer delays introduced by satellite links permeated the system. For a DSL system echo effects signal integrity and introduces unacceptable errors in data transmissions. An echo canceller synthesizers the echo path including the digital analog converter, the transmit filter, the hybrid circuit, the receiver filter, in the analogtodigital converter. An echo canceller can produce an echo replica with the same transmitting data but with reverse signed to cancel the real echo on the receive path.
 The topology of subscriber lines themselves may be used to minimize crosstalk between subscriber lines. Typically, telephone subscriber loops are organized in a binder with 10, 25, or 50 pairs each sharing a common physical or electrical shield in a cable. Due to capacitance and inductive coupling there's crosstalk between each twisted pair even though the pairs are well insulated for DC. The effective crosstalk is reduced by adapting different twist distances among different pairs in the binder group. Binder groups are also twisted such that no two groups are adjacent for long runs.
 The hybrid circuit which couples the modem to the subscriber line is also designed with noise reduction in mind. The hybrid is basically a bridge circuit which allows bidirectional communication on the subscriber line. When the bridge is balanced the spillover of noise from the modem's transmitted signal to the received signal is reduced. Balancing however requires an impedance match with the telephone subscriber loop which is never fully satisfied because the input impedance of the telephone loop varies from one loop to the next due to taps and temperature variations in the individual subscriber lines.
 Crosstalk noise comes from a adjacent telephone subscriber loops of the same or different types of transmission systems. Crosstalk is divided into what is known as near end cross talk (NEXT) and far end crosstalk (FEXT) depending on where the crosstalk is generated. NEXT is defined as crosstalk between subscriber lines in a binder coupled on one end with a common transceiver. FEXT is defined as a crosstalk affect between a receiving path and a transmitting path of the DSL transceivers on opposite ends of two different subscriber loops within the same twisted pair cable or binder. The FEXT noise at the receiver front end of a particular DSL transceiver is caused by signals transmitted by other transceivers at the opposite end of the twisted cable.
 What is needed is modem with improved capabilities for crosstalk cancellation.
 An apparatus for preceding multitone modulated transmissions of a plurality of communication channels over bundled subscriber lines. The apparatus includes a pseudo symbol controller and a precoder. The pseudo symbol controller detects a victim communication channel and an interferer communication channel having nonmatching symbol rates, and transforms the interferer channel into a pseudosymbol having both a substantially similar length as a corresponding symbol of the victim channel together with defined tonal characteristics expressed in terms of the interferer channel. The precoder assigns crosstalk coefficients to selected subchannels or tones of the pseudo symbol and precodes the victim channel with the pseudo symbol using the crosstalk coefficients; thereby generating a precoded victim symbol which exhibits reduced crosstalk between the selected victim and interferor.
 Associated means and methods are also disclosed.
 These and other features and advantages of the present invention will become more apparent to those skilled in the art from the following detailed description in conjunction with the appended drawings in which:

FIG. 1 shows a communication system with a pair of multimode multichannel modem line cards coupled to one another by a binder of subscriber lines between a public switched telephone network (PSTN) central office (CO) and remote sites. 
FIG. 2 is a detailed hardware block diagram of a precoder portion of one of the modem line cards shown inFIG. 1 . 
FIG. 3 is a detailed data structure view in the frequency domain of two communication channels with dissimilar symbol rates; 
FIG. 4 is a data structure view in the frequency domain of two communication channels shown inFIG. 3 including synchronization symbol portions thereof; 
FIGS. 5A5B are simplified transmission diagrams showing training and showtime phases respectively of precoding operation; 
FIG. 6 is a graph showing the crosstalk weighting assigned to a tone in the pseudo channel expressed in terms of its separation in frequency from a contributing tone in the disturber channel; 
FIGS. 7A7C are data structure diagrams showing various combinations of victim and disturber channel symbol rates and the combination of disturber channel symbols into corresponding pseudosymbols; 
FIGS. 8A8B are process flow diagrams for precoding during both training and showtime phases of operation respectively.  An apparatus and method for reducing interference over a common communication medium, wired or wireless is provided. The apparatus reduces interference from a number of sources using a common architecture which may be used to service a modem pool or discrete modems which share the common communication medium. Interference due to far end crosstalk (FEXT) may all be substantially reduced or cancelled by the apparatus. Additionally, channel characteristics for individual data channels across the common communication medium may be determined. The apparatus provides support for multiple modem protocols including XDSL protocols such as G.Lite, ADSL, VDSL, SDSL, MDSL, RADSL, and HDSL. The apparatus supports discrete multitone (DMT) line codes associated with XDSL communications. The apparatus may be implemented in hardware, firmware or software.

FIG. 1 shows a communication system with a pair of multimode multichannel modem line cards coupled to one another by a binder of subscriber lines between a public switched telephone network (PSTN) central office (CO) and a remote site. The system includes a CO 100 and a remote transceivers 156, 158, 160. The CO and remote line card are coupled to one another via a subscriber line binder 170 which includes individual subscriber lines 172, 174, 176.  Each of the subscriber line connections terminates on the CO end, in the frame room 102 of the CO. From this room connections are made for each subscriber line via splitters and hybrids to both a DSLAM 104 and to the voice band racks 106. The splitter shunts voice band communications to dedicated line cards, e.g. line card 112 or to a voice band modem pool (not shown). The splitter shunts higher frequency XDSL communications on the subscriber line to a selected line card, e.g. line card 120, within DSLAM 104. The line cards of the current invention are universal, meaning they can handle any current or evolving standard of XDSL and may be upgraded on the fly to handle new standards.
 Voice band call set up is controlled by a Telco switch matrix 114 such as SS7. This makes pointtopoint connections to other subscribers for voice band communications across the public switched telephone network 132. The XDSL communications may be processed by a universal line card such as line card 120. That line card includes a plurality of AFE's 134136 each capable of supporting a plurality of subscriber lines. The AFEs are coupled via a packet based bus 132 to the DSP 122. For downstream communications from the CO to the remote site, the DSP modulates the data for each communication channel, the AFE transforms the digital symbol packets assembled by the DSP and converts them to an analog signal which is output on the subscriber line associated with the respective channel. For upstream communications from the remote site to the CO the AFE each received channel is converted to a digitized data sample which is sent to the DSP. The DSP is capable of multiprotocol support for all subscriber lines to which the AFE's are coupled. Communications between AFE's, and DSP(s) may be packet based. The line card 120 is coupled to a backplane bus 116 which may be capable of offloading and transporting low latency XDSL traffic between other DSPs for load balancing. The backplane bus of the DSLAM also couples each line card to the Internet 130 via server 108. Each of the DSLAM line cards operates under the control of a DSLAM controller 110 which handles global provisioning, e.g. allocation of subscriber lines to AFE and DSP resources. The various components on the line card form a plurality of logical modems each handling upstream and downstream communications across corresponding subscriber lines. When an XDSL communication is established on a subscriber line, a specific channel identifier is allocated to that communication. That identifier is used in the above mentioned packet based embodiment to track each packet as it moves in an upstream or downstream direction between the AFE and DSP.
 In an alternate embodiment of the invention the termination at the remote site would be a single logical modem.
 The DSP 122 is shown with transmit path components 124, 126, 128, and receive path components 130. The precoder 126 operates during a training phase to establish coefficients for canceling Far end cross talk (FEXT). FEXT occurs as a result of transmissions to the remotes on subscriber lines 172 and 176 leaking into transmissions 194 from the CO on subscriber line 174 This leakage is represented by arrows 182,186 from subscriber lines 172, 176 respectively into the channel 194 received at modem 158. Near end crosstalk (NEXT) results from transmissions from the CO leaking over into reception at that same location. This leakage is represented by arrows 192,196 from subscriber lines 172, 176 respectively into the channel 194 received at the CO on subscriber line 174. SelfNEXT, a.k.a. echo, occurs on every subscriber line including line 174.

FIG. 2 is a detailed hardware block diagram of a precoder portion of one of the modem line cards shown inFIG. 1 . The transmit path components include framers and tone orderer 200, mapper 204, precoder 126, Inverse Discrete Fourier Transform components 224 and remaining transmit path components 226.  The precoder 126 modifies the transmit signal of a line to precompensate for the expected crosstalk from the other lines. For every line for which crosstalk should be canceled, the precoder modifies the modulation symbol (i.e. mapper output) on each of the tones by subtracting a linear combination of the modulation symbols (on the corresponding tone) of the disturbing lines. The factors in the linear combination are typically the crosstalk channel from the disturbing line divided by the direct channel of the victim line. In this invention, the precoder shows a number of additional blocks.
 The pseudosymbol controller 220 handles pseudo symbol detection, aggregation from the disturber channel and calculations related thereto including generation of the frequency content of the pseudosymbol. The pseudosymbol is a hypothetical symbol on the disturber line that has the same length as a symbol on the victim line. In the time domain it is obtained by considering a number of samples on the disturber line that coincides with the number of samples that are comprised in a symbol on the victim line. In the frequency domain, the pseudosymbol can be expressed as a combination of the frequency contents of the disturber symbols whose samples contribute to a particular pseudosymbol. The Tone selector 216 assembles the frequency contents of the relevant disturber symbols and output a tonebytone value for the pseudosymbol. Typically, multiple tones of the disturber symbol will contribute to a single tome of the pseudosymbol. The number of tones that contribute could be truncated to limit complexity. In that case, the toneselector 216 would determine which tones from the disturber/interferor symbols 214 to consider for a given tone of the pseudo symbol.
 The weighting of the various disturber tones that contrite to a tone of the pseudosymbol is determined by a configurable function 218. This function can be calculated (offline) for a given alignment of disturber and victim symbols. The output of the pseudosymbol controller is sent to the precoder 212. The precoder takes as input the crosstalk channel coefficients and the various direct channels 210 of the lines connected to the precoder.

FIGS. 5A5B are simplified transmission diagrams showing training and showtime phases respectively of precoding operation. Transmitter 500 is shown transmitting data across two bundled subscriber lines 502 to remote receivers, e.g. receivers 510 and 520. During the training phase of operation shown inFIG. 5A , the crosstalk from the disturber line 2 affects the noise level at the receiver end 510 of victim line 1. The resultant received signal 530 is passed back 532 to the transmitter during training to evaluate the crosstalk channel. The crosstalk is represented by a term H21*X2 in 530, where H21 is the crosstalk channel from lin2 into line 1 (i.e. the coupling from the transmitter 500 of line 2 to the receiver 510 of line 1). H1 is the “direct/victim channel”, i.e. the coupling from the transmitter of line 1 to the intended receiver 510 on line 1. The training allows the modems to estimate the value of H21.FIG. 5B illustrates the concept of precoding, or precompensating the crosstalk. In order to precompensate, a term −H21/H1*X2, reference 550, is added to the transmission on line 1. At the traversing the direct channel to the receiver of line 1, this term will have changed to become −H1*(H21/H1*X2)=−H21*X2. As such, it has the same value as the expected crosstalk from line 2, only with opposite sign. Therefore, the added term will cancel the physical crosstalk from line 2 into line 1 as exhibited in the resultant received signal 540, which exhibits no crosstalk.  Crosstalk canceling in DSL is generally conceived as a frequency domain process. It can be accomplished for bundled subscriber lines coupled on one end to a modem pool, and to discrete modems on the opposing end, a.k.a. vectoring, or on a bundle of subscriber lines coupled at both opposing ends to associated modem pools, a.k.a. multipleinput multiple output (MIMO). By synchronizing symbols on different lines, crosstalk canceling can be performed using a single tap per tone (and per crosstalker). When the system consists of a mixed deployment of VDSL2 systems operating at 4 kHz and 8 kHz symbols rates, it is no longer possible to align the systems and achieve the required orthogonality. As a result, crosstalk can no longer be canceled per tone. In the following embodiment of the invention a frequencydomain approach to canceling crosstalk from a system with 8 kHz symbol rate into a system with a 4 kHz symbol rate is disclosed.
 For crosstalk channel estimation, content and location of the sync symbols (which are used to obtain the crosstalk channels) is modified such that 4 kHz and 8 kHz lines share a common sync symbol, even though the data rates on the lines are different. This involves doubling the length of the sync symbol on the 8 kHz line and limiting the frequencycontent of the sync symbol to even tones only on the 4 kHz line.
 For crosstalk canceling, the even tones can be dealt with using a single operation per tone, while the odd tones require information from multiple tones to be combined. Various levels of implementation complexity can be considered by limiting the number of odd tones that is considered during the canceling process. The slicing process of sync symbols can operate in much the same way as is done for lines with equal rates. The canceling process will be handled externally by a dedicated chip.
 Crosstalk canceling in DSL facilitated by the fact that canceling can be performed as a single operation per tone (and per crosstalker) when DMT symbols across different lines are aligned and synchronized to a common clock. A similar operation in the time domain would require significantly more operations. In general, the basic requirements to enable simple frequency domain cancellation are:

 symbols across different lines need to be synchronized
 all symbols need to be the same length
 sync symbols on different lines need to be synchronized to allow for easy estimation of the crosstalk channel
VDSL2 has defined a number of different profiles. While most profiles specify a 4 kHz symbol rate, profile 30 a has a symbol rate equal to 8 kHz. This makes it fundamentally impossible to comply with the basic requirements outlined above. As deployments go, it is not always possible to assume that only profiles with 4 kHz tone spacing (resp. 8 kHz tone spacing) will be present. The VDSL2 system may switch profiles as a function of loop length and other factors. As such, a mixed deployment containing both systems at 4 Hz symbol rate and systems at 8 kHz symbol rate can occur in practice.
 Crosstalk canceling between systems with different symbol rates (specifically 4 kHz and 8 kHz symbol rates is disclosed, with a particular example being 8 kHz interferors and 4 kHz victim/direct channels.
 Crosstalk channel estimation relies on the data gathered from sync symbols. All sync symbols in a vectored system are aligned and the sign is modulated with an orthogonal sequence. When this is done, information about the crosstalk channels can be derived from the slicer errors of consecutive sync symbols. This information can be recovered by appropriate processing. Synchronization of sync symbols can currently not be achieved when systems with different symbol rates are considered. VDSL2 systems insert a sync symbol after each 256 data symbols. The sync symbols of systems operating at 4 kHz and 8 kHz symbol rates will not coincide as illustrated in
FIG. 4 , specifically 4 kHz channel 400 and 8 kHz channel 410. While there are other ways of performing crosstalk channel estimation, the method that relies on synchronized sync symbols and orthogonal sequences is by far the most convenient. This method can still be used provided that we make some modifications to the content and the location of the sync symbols on the 8 kHz lines. Specifically as shown in modified 8 kHz symbol stream 420 inFIG. 4 , instead of sending a sync symbol every 256 data symbols, those lines will send two backtoback sync symbols every 512 symbols. In addition, the second symbol in each pair of backtoback symbols is phase rotated to assure a smooth transition between the two sync symbols. As a result, the two contiguous sync symbols look exactly the same as a sync symbol sent on the 4 kHz line, with the only difference that only the even tones, on a 4 kHz tone grid, contain power. Also, with this embodiment of the invention, the number of sync symbols and therefore the total overhead associated with sync symbols remains the same. When the symbols are aligned this way, sync symbols of a 4 kHz line are orthogonal to the sync symbols of an 8 kHz line and vice versa. In this embodiment of the invention the slicer errors on a give tone are only affected by the signal sent on that tone on other lines. The same method can be maintained for estimating crosstalk channels. Note that this will only produce the crosstalk channel values at multiples of 8 kHz. However, the value of the crosstalk channel on the intermediate tones can be estimated through interpolation.  Although the symbols of a 4 kHz system and an 8 kHz systems cannot be aligned exactly, we'll assume that the 8 kHz symbols and 4 kHz symbols are aligned such that each 4 kHz symbol coincides with exactly two 8 kHz symbols. This is illustrated in
FIG. 3 where the 4 kHz symbol 300 with a cyclic prefix 302 and 8 k symbols 314 and 320 with cyclic prefixes 316 and 322 respectively are shown. With this alignment, two and only two 8 kHz symbols 310 will contribute to the crosstalk experienced by the 4 kHz symbol. The first step is to calculate the crosstalk in terms of the frequencydomain content of those two 8 kHz symbols. InFIG. 3 , the following notation is used: 
 CE4 is the cyclic extension of the 4 kHz symbol
 CE8 is the cyclic extension of the 8 kHz symbol. This value is exactly half of CE4, otherwise synchronization would not be possible.
The receiver will determine where to align the FFT window and which 2N samples to send to the FFT input. InFIG. 3 , we have indicated the FFT alignment 348 with an offset delta relative to the start of the symbol (counting from the first sample of the cyclic extension). While this alignment typically is a receiverproprietary decision, the most advantageous choice for purposes of crosstalk canceling is delta=CE8. For this value, each of the two 8 kHz symbols contributes exactly N samples to the FFT input, as illustrated inFIG. 3 pseudosymbol 328. For the remainder of the calculations, we'll assume that this alignment is used.
Note that although each 8 kHz symbol inFIG. 3 contributes exactly N samples, the N samples from the second symbol start at the first sample of the cyclic prefix. These samples are not protected by the cyclic extension and may be subject to intersymbol interference. This will limit the amount of cancelation that can be achieved in practice. Nevertheless, the resulting gains in SNR are still significant however. Note also, that if the 8 kHz system uses a larger spectrum than the 4 kHz system, the design of the 4 kHz system should be done such that aliasing of these higher frequencies does not affect the spectrum used by the 4 kHz system. Starting from the alignment shown inFIG. 3 , we can calculate how the two 8 kHz symbols appear at the output of the 2Npoint FFT.
We will denote the samples of the two 8 kHz symbols as y_{n }and z_{n }respectively, such that:

$\begin{array}{cc}{y}_{n}=\sum _{k=0}^{N1}\ue89e{Y}_{k}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}\ue89e\text{}\ue89e{z}_{n}=\sum _{k=0}^{N1}\ue89e{Z}_{k}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}& \left(1\right)\end{array}$  The samples that go into the 2Npoint FFT are given by:

FFT_{2N}[y_{0 }. . . y_{NCE8 }. . . z_{N1 }z_{0 }. . . z_{NCE81}] (2)  This is equal to:

$\begin{array}{cc}F\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eF\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{T}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}\left[\begin{array}{cccc}{y}_{0}& \dots & {y}_{N1}& \stackrel{\stackrel{N}{\uf612}}{0\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e0}\end{array}\right]+F\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eF\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{T}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}\left[\begin{array}{ccccccc}\stackrel{\stackrel{N}{\uf612}}{0\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e0}& {z}_{N\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8}& \dots & {z}_{N1}& {z}_{0}& \dots & {z}_{N\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e81}\end{array}\right]& \left(3\right)\end{array}$  The first term in (3) can be calculated as:

$\begin{array}{cc}\begin{array}{c}{\stackrel{~}{Y}}_{k}=F\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eF\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{T}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}\left[{y}_{0}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e{y}_{N1}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\stackrel{\stackrel{N}{\uf612}}{0\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e0}\right]\\ =\sum _{n=0}^{N1}\ue89e\left(\sum _{{k}^{\prime}=0}^{N1}\ue89e{Y}_{k}\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{N}}\right)\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}\\ =\frac{1}{N}\ue89e\sum _{{k}^{\prime}=0}^{N1}\ue89e{Y}_{{k}^{\prime}}\ue89e\frac{1{\left(1\right)}^{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}k}}{1{\uf74d}^{j\ue89e\frac{\pi \ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}k\right)}{N}}}\end{array}& \left(4\right)\end{array}$  Likewise, the second term in (3) can be calculated as:

$\begin{array}{cc}\begin{array}{c}{\stackrel{~}{Z}}_{k}=\ue89e{\mathrm{FFT}}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}\left[\stackrel{\stackrel{N}{\uf612}}{0\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e0}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e{z}_{N\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e{z}_{N1}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e{z}_{0}\ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e\dots \ue89e\phantom{\rule{0.8em}{0.8ex}}\ue89e{z}_{N\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e81}\right]\\ =\ue89e\sum _{n=N}^{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN1}\ue89e\left(\sum _{{k}^{\prime}=0}^{N1}\ue89e{Z}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(n\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{N}}\right)\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\\ =\ue89e\frac{1}{N}\ue89e{\uf74d}^{j\ue89e\frac{\pi \xb7\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\xb7k}{N}}\ue89e\sum _{{k}^{\prime}=0}^{N1}\ue89e{Z}_{{k}^{\prime}}\ue89e\frac{1{\left(1\right)}^{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}k}}{1{\uf74d}^{j\ue89e\frac{\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e{k}^{\prime}k\right)}{N}}}\xb7{\uf74d}^{j\ue89e\frac{\pi \ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}k\right)\ue89e\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8}{N}}\end{array}& \left(5\right)\end{array}$  From (4) and (5), we note that there is a crucial difference between the way even tones and odd tones of the 4 kHz symbol are affected by the two 8 kHz symbols. For even tones, the contribution of the two 8 kHz symbols reduces to:

$\begin{array}{cc}\begin{array}{c}{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el}={\stackrel{~}{Y}}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el}+{\stackrel{~}{Z}}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el}\\ ={Y}_{l}+{\uf74d}^{j\ue89e\frac{\pi \xb7\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\xb72\ue89el}{N}}\ue89e{Z}_{l}\end{array}& \left(6\right)\end{array}$  In other words: the even tones are only affected by a single tone of the 8 kHz symbols and the impact is a linear combination of the constellation points modulated on that tone during each of the two 8 kHz symbols. This makes the cancellation of the crosstalk on the even tones only twice as complex as the cancellation of the crosstalk from a 4 kHz symbol. Instead of a single tap canceller, a twotap canceller is needed for each of the even tones.
For odd tones, the situation is more complex. The contribution of the two 8 kHz symbols becomes: 
$\begin{array}{cc}\begin{array}{c}{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el+1}=\ue89e{\stackrel{~}{Y}}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el+1}+{\stackrel{~}{Z}}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el+1}\\ =\ue89e\frac{1}{N}\ue89e\sum _{{k}^{\prime}=0}^{N1}\ue89e{Y}_{{k}^{\prime}}\ue89e\frac{2}{1{\uf74d}^{j\ue89e\frac{\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\left({k}^{\prime}l\right)1\right)}{N}}}\\ \ue89e\frac{1}{N}\ue89e{\uf74d}^{j\ue89e\frac{\pi \xb7\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\xb7\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89el+1\right)}{N}}\\ \ue89e\sum _{{k}^{\prime}=0}^{N1}\ue89e{Z}_{{k}^{\prime}}\ue89e\frac{2}{1{\uf74d}^{j\ue89e\frac{\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\left({k}^{\prime}l\right)1\right)}{N}}}\xb7{\uf74d}^{j\ue89e\frac{\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\left({k}^{\prime}l\right)1\right)}{N}}\end{array}& \left(7\right)\end{array}$  Potentially, all tones of the 8 kHz symbols will contribute to a single odd tone of the 4 kHz symbol. This makes full cancellation of the crosstalk on the even tones extremely complex.
 Note however that not all 8 kHz tones in (7) contribute equally to the final output X_{2l+1}. In fact, the weighting function peaks sharply around k′=l, as illustrated in
FIG. 6 . This means that the crosstalk on the even tones can still be cancelled partially by only considering a limited number of terms in the sums in (5). For a low number of terms the complexity will still be relatively low. The higher the number of terms, the higher the complexity, but the higher the gain that can be achieved.  Canceling of 8 kHz Crosstalk into 4 kHz
Crosstalk cancelling of the 8 kHz signal into a 4 kHz symbol can be achieved by precoding, similarly to what is done in the case when all symbols have the same symbol rate. If the symbol modulated on tone k of the 4 kHz symbol is denoted as A_{k}, the precoding is achieved by adding the following terms to A_{k}: 
$\begin{array}{cc}\phantom{\rule{4.4em}{4.4ex}}\ue89e{A}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\to {A}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\frac{{H}_{\mathrm{mn}}\ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\right)}{{H}_{\mathrm{mm}}\ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\right)}\ue89e\left({Y}_{k}+{\uf74d}^{j\ue89e\frac{\pi \xb7\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\xb72\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}{N}}\ue89e{Z}_{k}\right)\ue89e\text{}\ue89e{A}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1}\to {A}_{2\ue89ek+1}\frac{{H}_{\mathrm{mn}}\ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{{H}_{\mathrm{mm}}\ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}\ue89e\frac{1}{N}\ue89e\sum _{{k}^{\prime}=0}^{N1}\ue89e\left(\begin{array}{c}{Y}_{{k}^{\prime}}{\uf74d}^{j\ue89e\frac{\pi \xb7\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\xb7\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{N}}\ue89e{Z}_{{k}^{\prime}}\xb7\\ {\uf74d}^{j\ue89e\frac{\pi \ue8a0\left(2\ue89e\left({k}^{\prime}k\right)1\right)\ue89e\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8}{N}}\end{array}\right)\ue89e\frac{2}{1{\uf74d}^{j\ue89e\frac{\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\left({k}^{\prime}k\right)1\right)}{N}}}& \left(8\right)\end{array}$  In (8), H_{mm }denotes the transfer function of the direct channel, while H_{mn }denotes the crosstalk transfer function between line “n” and “m”. The 4 kHz system is deployed on line “m”, while the 8 kHz system is deployed on line “n”. For the odd tones, the number of terms in the precoding can be limited to achieve either the desired level of crosstalk canceling or the desired complexity.
 A simulation follows that implements the crosstalk canceling of an 8 kHz system into a 4 kHz system using the methodology outlined above. Both systems were connected to 200 meter loops of gauge 26 awg. To model the impulse response of the crosstalk channel, we took the 99% worstcase coupling and further assumed that the phase behavior of the crosstalk channel was identical to the phase behavior of the direct channel. In other words, the transfer function of the crosstalk channel in the frequency domain was given by:

H _{Xtalk}(f)=K·f·√{square root over (L)}·H(f), (9)  K=1.5940.10^{−10 }
 L is the loop length (200 meters)
 H(f) is the transfer function of the direct channel
 For the purposes of this analysis, tones 100 to 2000 (on the 4 kHz tone grid) were loaded with random 4 QAM symbols. The crosstalk from the 8 kHz systems occurs on the same tones. During the simulation, DMT symbols were generated for both a 4 kHz symbol rate and an 8 kHz symbol rate. The samples corresponding to the 4 kHz symbol are sent through the direct channel by performing convolution with the impulse response of the direct channel. The samples corresponding to the 8 kHz symbol are sent through the crosstalk channel by performing convolution with the impulse response of the crosstalk channel. The resulting output samples are added to make up the total receive signal consisting of useful signal and crosstalk. The receiver performs the symbol alignment as shown in
FIG. 3 and calculates the SNR.
The same is done for a 4 kHz system with precoding as described in formula (8). The SNR for the precoded case is calculated and compared to the SNR without precoding. 
FIG. 7A7B show pseudosymbols for even and odd tones respectively of a 8 kHz disturber and a 4 k victim/direct channel. The interferor symbols 740 and 742 which comprise the 8 kHz pseudo symbol are shown along with the symbol 744 for the victim/direct 4 kHz channel. For even tones, the precoding can be done with a twotap precoder per tone. As shown in (8), the crosstalk compensation that is added to an evennumbered tone at the transmitter is a linear combination of the constellation points that are modulated on that frequency during each of the two 8 kHz symbols. The SNR with and without crosstalk canceling is shown in The SNR improves significantly after crosstalk canceling. The typical improvement is between 30 and 40 dB for all of the tones. This is achieved at a computational complexity that is about double the complexity that is required for canceling crosstalk from a 4 kHz system.  For oddnumbered tones, the precoding becomes more complex. Theoretically, each oddnumbered tone is affected to some extent by all of the 8 kHz tones. The SNR improvement that can be achieved will depend on the complexity on is willing to tolerate. Note that in the limit, one may even decide to only cancel the crosstalk on the evennumbered tones and not attempt any improvement on the odd tones.
The gain depends to a large extent on the number of terms that is included in the precoding. shows the SNR average improvement over the tones for the different truncations. Even for as low as three tones, an improvement of more than 8 dB can be achieved. As more terms are included, the performance gain goes up slowly. The full gain of about 40 dB is only achieved when all terms are taken into account.
The following techniques can help in maintaining a feasible level complexity for crosstalk canceling: 
 synchronize the symbols of the different lines such that each 4 kHz symbol overlaps with exactly two 8 kHz symbols
 modify the location of the sync symbols on the 8 kHz line to coincide with the sync symbols of the 4 kHz line. Also modify the symbols on the 8 kHz line so that they appear identical on both lines.
 Canceling of the crosstalk from an 8 kHz symbol into the evennumbered tones of a 4 kHz symbol requires a twotap operation that represents roughly twice the complexity that would be required for the crosstalk cancelation of a 4 kHz system.
 Canceling of the crosstalk from an 8 kHz symbol into the oddnumbered tones of a 4 kHz symbol potentially requires a linear combination of the constellation points sent on all loaded tones of the 8 kHz symbols. Because of the nature of the weighting function, the number of terms can be limited while still achieving moderately good SNR improvements.
 The 4 kHz symbols and 8 kHz symbols are aligned such that each 4 kHz symbol overlaps with exactly two 8 kHz symbols. The same assumption will be made in this contribution. The alignment of symbols is illustrated in
FIG. 3 and further inFIG. 7C where the 4 kHz disturber/interferor pseudosymbol 760 and the pair of 8 kHz victims 762 and 764 are shown. Again, the value “N” denotes the number of tones in the 4 kHz symbol. For the reverse case of 8 kHz crosstalk into a 4 kHz symbol discussed in above, we assumed a specific alignment of the FFT window at the 4 kHz receiver relative to the position of the 8 kHz symbols. The FFT window was offset by CE8 samples relative to the start of the 4 kHz symbol. This was done to ensure that the 8 kHz symbols affecting the 4 kHz symbol each contributed exactly N samples to the 2N input samples of the (4 kHz) FFT. This allowed for a simplified treatment of the crosstalk cancelation.  For this case, 4 kHz crosstalk into 8 kHz receiver, no such constraint on the FFT alignment is needed. The N input samples into the 8 kHz FFT can be taken freely from any starting point within the N+CE8 samples of the 8 kHz DMT symbol. For any chosen alignment, the crosstalk samples will belong to a single 4 kHz DMT symbol. In FIG. 3the offset of the N input samples relative to the start of the 8 kHz DMT symbol is denoted by delta. The constellation points that are modulated on the tones of the 4 kHz symbols a X_{k}, with k=0, . . . , 2N1, with:

X _{2Nk} =X* _{k } (1) 
X_{0}=0=X_{N }  The 2N samples of the body of the 4 kHz symbol are then given by:

$\begin{array}{cc}{x}_{n}=\sum _{k=0}^{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}& \left(2\right)\end{array}$  Using (1), this can be rewritten as:

$\begin{array}{cc}\begin{array}{c}{x}_{n}=\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}+\sum _{{k}^{\prime}=N}^{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\\ =\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}+\sum _{{k}^{\prime}=1}^{N}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue8a0\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN{k}^{\prime}\right)\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\\ =\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}+\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}^{*}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\end{array}& \left(3\right)\end{array}$  Since each 4 kHz symbol affects two 8 kHz symbols, we'll look separately at the “first” and “second” 8 kHz symbol that overlaps with the 4 kHz symbol For a specific alignment as shown in the samples of the 4 kHz symbol that contribute to FFT input of the first 8 kHz symbol are given by:

s _{n} ^{(1)} =x _{2NCE4+Δ} x _{2NCE4+Δ+1 } . . . x _{2N1 } x _{0 } x _{1 } . . . x _{N1CE4+Δ}] (4)  Likewise, the samples of the 4 kHz symbol that contribute to FFT input of the second 8 kHz symbol are given by:

s _{n} ^{(2)} =[x _{NCE8+Δ} x _{NCE8+Δ+1 } . . . x _{2N1CE8+Δ}] (5)  The intention is to cancel the crosstalk coming from the 4 kHz symbol into the 8 kHz symbols in the frequency domain. For this purpose, consider the N “crosstalk samples” given by either (4) or (5) as samples coming from an equivalent 8 kHz symbol. Thus a set of constellation points such that the Npoint IFFT of these constellation points generates the same sample values as (4) and (5) is required. A calculation of the frequency representation of this equivalent 8 kHz symbol and express it in terms of the constellation points X_{k }of the 4 kHz symbol is required. Finding the frequency representation of the samples s_{n} ^{(1) }can be done by simply taking the FFT of (4):

$\begin{array}{cc}\begin{array}{c}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(1\right)}\right)=\sum _{n=0}^{N1}\ue89e{s}_{n}^{\left(1\right)}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}\\ =\sum _{n=0}^{N1}\ue89e{x}_{n\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4+\Delta}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}\\ =\sum _{n=0}^{N1}\ue89e\left(\begin{array}{c}\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(n\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4+\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}+\\ \sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}^{*}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(n\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4+\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\end{array}\right)\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\mathrm{kn}}{N}}\end{array}& \left(6\right)\end{array}$  Inverting the order of the sums over n and k′, we get:

$\begin{array}{cc}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(1\right)}\right)=\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\right)\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}+\sum _{{k}^{\prime}=0}^{N1}\ue89e{X}_{{k}^{\prime}}^{*}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}+2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\right)\ue89en}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}& \left(7\right)\end{array}$  Separating out the even and odd terms in the sum over k′, we get:

$\begin{array}{cc}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(1\right)}\right)=\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}k\right)\ue89en}{N}}+{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)\ue89en}{N}}+\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}}^{*}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}+k\right)\ue89en}{N}}+{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}^{*}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)\ue89en}{N}}& \left(8\right)\end{array}$  This can be further simplified by using the following identities:

$\begin{array}{cc}\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}k\right)\ue89en}{N}}=N\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{\delta}_{{\mathrm{kk}}^{\prime}}\ue89e\text{}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}+k\right)\ue89en}{N}}=N\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{\delta}_{k\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e0}\ue89e{\delta}_{{k}^{\prime}\ue89e0}\ue89e\text{}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)\ue89en}{N}}=\frac{2}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left({k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}\ue89e\text{}\ue89e\sum _{n=0}^{N1}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)\ue89en}{N}}=\frac{2}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}& \left(9\right)\end{array}$ 
$\begin{array}{cc}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(1\right)}\right)=N\xb7{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}+2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}+N\xb7{X}_{0}+2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}^{*}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}& \left(10\right)\end{array}$  The third term in (10) can be dropped if we assume that no energy is sent at DC (which is typically the case). Also, the contribution of the fourth term is very small compared to the other remaining terms. This means that for practical purposes, (10) reduces to:

$\begin{array}{cc}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(1\right)}\right)=N\xb7{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}+2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}& \left(11\right)\end{array}$  Formula (11) gives us the frequency representation of the N samples (4). It's straightforward to see that the IFFT of these frequencydomain points will results in the N samples (4), since the expression (11) is just the FFT of those samples. Note that the frequency domain representation depends only on the constellation points of the 4 kHz crosstalk symbol and the alignment of the FFT input samples relative to the symbol boundaries.
The frequency representation of the samples s_{n} ^{(2) }(see (5)) is done in much the same way. After a similar calculation, we find: 
$\begin{array}{cc}{\mathrm{FFT}}_{k}\ue8a0\left({s}^{\left(2\right)}\right)=N\xb7{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\ue8a0\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{N}}2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\ue89e\sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}& \left(12\right)\end{array}$  Precoding is done differently for “evennumbered” and “oddnumbered” 8 kHz DMT symbols. As can be seen form (11) and (12), the 4 kHz crosstalk symbol will affect the first 8 kHz symbol that overlaps with the 4 kHz symbol differently than the second 8 kHz symbol that overlaps with the same 4 kHz symbol. The constellation point modulated on tone k of the 8 kHz symbol (i.e. the tone with frequency k*8.625 kHz) is denoted as A_{k}, k=0, . . . , N/21. For the evennumbered symbols, we use the result (11). This leads to the following precoding:

$\begin{array}{cc}{A}_{k}\to {A}_{k}\frac{{H}_{\mathrm{mn}}\ue8a0\left(k\right)}{{H}_{\mathrm{mm}}\ue8a0\left(k\right)}\ue89e\left(\begin{array}{c}N\xb7{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}+2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\\ \sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e4\Delta \right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}\end{array}\right)& \left(13\right)\end{array}$  For the oddnumbered symbols, we use the result (12). This leads to the following preceding:

$\begin{array}{cc}{A}_{k}\to {A}_{k}\frac{{H}_{\mathrm{mn}}\ue8a0\left(k\right)}{{H}_{\mathrm{mm}}\ue8a0\left(k\right)}\ue89e\left(\begin{array}{c}N\xb7{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek}\ue89e{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek\ue8a0\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{N}}2\xb7{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}\\ \sum _{{k}^{\prime}=0}^{N/21}\ue89e{X}_{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}+1}\ue89e\frac{{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}\ue8a0\left(\Delta \mathrm{CE}\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e8\right)}{N}}}{1{\uf74d}^{j\ue89e\frac{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\pi \ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e\left(2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89e{k}^{\prime}2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89ek+1\right)}{2\ue89e\phantom{\rule{0.3em}{0.3ex}}\ue89eN}}}\end{array}\right)& \left(14\right)\end{array}$  The crosstalk affecting tone k of an 8 kHz symbol has contributions from a single even tone of the 4 kHz symbol (the tone that coincides with the frequency of the 8 kHz tone), as well as contributions from all odd (4 kHz) tones. Similarly to the case of 8 kHz crosstalk into 4 kHz symbol, the contributions of the odd tones are not evenly weighted.
 Both the crosstalker (the 4 kHz system) and the victim system (the 8 kHz system) are connected to 26 awggauge loops of 50 meter length. To model the frequency response of the crosstalk channel, a 99% worstcase FEXT coupling was used, further assuming that the phase behavior of the crosstalk channel was identical to the phase behavior of the direct channel. For the purposes of the simulation, tones 50 to 1000 (on the 8 kHz grid) were used for transmission. The crosstalk canceller only works on these tones. Transmission is modeled as 4 QAM on all active subcarriers. During the simulation, DMT symbols are generated for both a 4 kHz symbol rate and an 8 kHz symbol rate. The samples corresponding to the 8 kHz symbol are sent through the direct channel by performing convolution with the impulse response of the direct channel. The samples corresponding to the 4 kHz symbol are sent through the crosstalk channel by performing convolution with the impulse response of the crosstalk channel. The resulting output samples of the direct channel and the crosstalk channel are added to make up the total receive signal consisting of useful signal and crosstalk. Note that the DMT symbols for the different symbol rates are aligned as shown in
FIG. 3 The receiver performs the symbol alignment and calculates the SNR.  The results of the crosstalk canceling obviously depend on the number of terms that is included in the preceding (see formulas (13) and (14)). To achieve the maximum gain, up to N/2 terms should be included in the preceding of a single tone. This is not a practical solution in implementation, but it establishes the upper bound of what can be achieved. The results depend strongly on the number of terms in the precoding structure. Note also that in this case, there is no difference between odd and even tones. The implementation with minimum complexity is obtained when only the first term in (13) or (14) is used in the preceding (denoted as “no odd tones” in This ignores the contributions of all odd 4 kHz tones to the crosstalk on the 8 kHz tones. This implementation has the same complexity as the precoder for a set of systems that all use the same tone spacing. However, the performance improvement that can be realized in this case is very limited. The performance goes up gradually as more terms are added to the precoding. Crosstalk canceling is possible, but requires a multitap preceding per tone. Each of the 8 kHz tones receives a crosstalk contribution from all 4 kHz tones, not just the tone with the same frequency. Because the crosstalk contribution of each of the 4 kHz tones decreases rapidly, it is feasible to limit the number of terms in the precoding. This allows for various performance vs. complexity tradeoffs.

FIGS. 8A8B are process flow diagrams for preceding bundled subscriber lines during both training and showtime phases of operation respectively. 
FIG. 8A shows training initiated in process 800 in which a new line(s) is initialized. Next in process 802804 the id of the next interferor/disturber and the next direct/victim channel is determined respectively. Then in decision process 806 a determination is made as to whether the associated symbol rates for victim and interferor channels match. If they do then control passes to process 816 in which crosstalk coefficients are assigned to the tones of the interferer. The higher the coefficient, the greater the crosstalk coupling factor. Alternately, if symbol rates do not match control passes to processes 808812. In process 808 a pseudosymbol is assembled from the interferer channel. The pseudosymbol matches the victim channel's sample number and symbol rate. Next in process 810 a closed form function or expression is generated which expresses frequency content of the pseudo symbol as a function of disturber/interferor symbols which comprise the pseudosymbol. Next in process 812 the frequency content of the pseudo symbol is determined based on the actual complex values of the contributing tones from the disturber channel and the weighting factors attached thereto as shown inFIG. 6 . Then in process 816 crosstalk coefficients are assigned to the tones of the pseudosymbol. The transmitted victim channel is precoded in process 820 to reduce crosstalk from the interferor/disturber channel. Precoding typically includes contributions from more than one interferor channel but can be handled one interferer at a time, depending on the training protocol. Next, slicer error indicia is received from the remote modem receiving the precoded victim channel. If slicer error is below an acceptable threshold then in decision process 824 crosstalk has been reduced to an acceptable level after which control passes to process 826. In process 826 all accumulated crosstalk cancellation information, functions, coefficients and weighting factors for the selected victim and disturber(s) channels is stored. Then in decision processes 828 and 830 further processing of the next victim channel or next interferor for the current victim channel is initiated. Next in process 832 a determination may be made as to the major interferors for each victim channel and limited processing resources may be apportioned accordingly. If such a determination is made, then in process 834 precode taps settings for each direct/victim channel are stored for the top interferors established in process 832. Finally, in decision process 836, a determination is made as to whether training is complete of if a new channel needs initialization. If there is a new channel requiring training then control returns to process 800. 
FIG. 8B shows the processes associated with showtime operation of the bundled subscriber lines, and specifically preceding thereof. Showtime is initiated in process 850. In processes 852854 the id of the next direct/victim channel and the first/next interferor/disturber channel is determined respectively. Then in decision process 856 a determination is made as to whether the associated symbol rates for victim and interferer channels match. If they do then control passes to process 866 in which crosstalk coefficients are assigned to the tones of the interferor. Alternately, if symbol rates do not match control passes to processes 858862. In process 858 a pseudosymbol is assembled from the interferer channel. The pseudosymbol matches the victim channel's sample number and symbol rate. Next in process 860 the closed form function or expression established during the training phase is loaded along with the assigned pseudochannel weights discussed above. Next in process 862 the frequency content of the pseudo symbol is determined based on the actual complex values of the contributing tones from the disturber channel and the weighting factors attached thereto as shown inFIG. 6 . Then in process 866 crosstalk coefficients determined in the training phase are assigned to the tones of the pseudosymbol. Next in decision process 870 a determination is made as to whether any further interferors for the subject channel remain. If there are then control returns to process 854 for further precode processing of the aggregate precode factor to be applied to the subject victim/direct channel. If precode calculation for all interferors is complete, then control passes to process 872 in which precoding of the direct/victim channel is completed. Next in decision process 880 a determination is made as to whether all victim/channels have been processed in the current symbol interval. If not control returns to process 852. If processing of all channels is complete, then control passes to process 886 for the showtime processing of all channels in the next symbol interval.  The foregoing description of a preferred embodiment of the invention has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise forms disclosed. Obviously many modifications and variations will be apparent to practitioners skilled in this art. It is intended that the scope of the invention be defined by the following claims and their equivalents.
Claims (19)
Priority Applications (3)
Application Number  Priority Date  Filing Date  Title 

US9135808P true  20080823  20080823  
US15144109P true  20090210  20090210  
US12/583,694 US8427933B2 (en)  20080823  20090824  Method and apparatus for DMT crosstalk cancellation 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

US12/583,694 US8427933B2 (en)  20080823  20090824  Method and apparatus for DMT crosstalk cancellation 
Publications (2)
Publication Number  Publication Date 

US20100046738A1 true US20100046738A1 (en)  20100225 
US8427933B2 US8427933B2 (en)  20130423 
Family
ID=41696410
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

US12/583,694 Active 20320130 US8427933B2 (en)  20080823  20090824  Method and apparatus for DMT crosstalk cancellation 
Country Status (4)
Country  Link 

US (1)  US8427933B2 (en) 
EP (1)  EP2319191A4 (en) 
KR (1)  KR101600332B1 (en) 
WO (1)  WO2010024884A1 (en) 
Cited By (11)
Publication number  Priority date  Publication date  Assignee  Title 

US20100034374A1 (en) *  20061020  20100211  Oehman Hans  Vectored digital subscriber line system having modular vectoring arrangements 
US20110235759A1 (en) *  20100323  20110929  Ikanos Communications, Inc.  Systems and Methods for Implementing a MultiSensor Receiver in a DSM3 Environment 
US20120026926A1 (en) *  20100531  20120202  Rudi Frenzel  Low power mode for vectored data transmission 
US20120114063A1 (en) *  20090827  20120510  Huawei Technologies Co., Ltd.  Method, apparatus and system for eliminating aliasing noise in multicarrier modulation system 
US20140056312A1 (en) *  20120723  20140227  Lantiq Deutschland Gmbh  Spectrum Management and Timing Optimization Over Multiple Distribution Points 
US20140119419A1 (en) *  20121030  20140501  Lantiq Deutschland Gmbh  Channel Estimation 
US20140314071A1 (en) *  20111231  20141023  Huawei Technologies Co., Ltd.  Method, Apparatus and System for MultiCarrier OFDM Duplex Transmission 
CN105453501A (en) *  20140724  20160330  华为技术有限公司  Crosstalk estimation method, device, and system 
US20160248475A1 (en) *  20150224  20160825  Lantiq BeteiligungsGmbH & Co. KG  Crosstalk mitigation 
EP3255807A4 (en) *  20150429  20171220  Huawei Technologies Co., Ltd.  Method, apparatus and system for sending signal 
CN109728837A (en) *  20171030  20190507  华为技术有限公司  A kind of method, apparatus and system for offsetting crosstalk signal 
Families Citing this family (1)
Publication number  Priority date  Publication date  Assignee  Title 

JP6124362B2 (en) *  20111102  20170510  マーベル ワールド トレード リミテッド  Method and apparatus for automatically detecting the physical layer (PHY) mode of a data unit in a wireless local area network (WLAN) 
Citations (8)
Publication number  Priority date  Publication date  Assignee  Title 

US20020027985A1 (en) *  20000612  20020307  Farrokh RashidFarrokhi  Parallel processing for multipleinput, multipleoutput, DSL systems 
US20020042836A1 (en) *  20000407  20020411  Mallory Tracy D.  Method of enhancing network transmission on a priorityenabled framebased communications network 
US20060259535A1 (en) *  20050510  20061116  Adaptive Spectrum And Signal Alignment, Inc.  Tonal rotors 
US20070230548A1 (en) *  20060317  20071004  Broadcom Corporation  Digital subscriber line noise mitigation techniques, and applications thereof 
US7400710B2 (en) *  20040226  20080715  Conexant, Inc.  MIMO dynamic PSD allocation for DSL networks 
US20080198909A1 (en) *  20030908  20080821  Michail Konstantinos Tsatsanis  Efficient multiple input multiple output signal processing method and apparatus 
US20090059780A1 (en) *  20070831  20090305  De Lind Van Wijngaarden Adriaan J  Method and apparatus for selftuning precoder 
US20090116582A1 (en) *  20071102  20090507  Alexei Ashikhmin  Interpolation method and apparatus for increasing efficiency of crosstalk estimation 
Family Cites Families (4)
Publication number  Priority date  Publication date  Assignee  Title 

US7072414B1 (en) *  19990907  20060704  The Aerospace Corporation  Gaussian minimum shift keying (GMSK) precoding communication method 
US6990196B2 (en) *  20010206  20060124  The Board Of Trustees Of The Leland Stanford Junior University  Crosstalk identification in xDSL systems 
US8073135B2 (en) *  20050510  20111206  Adaptive Spectrum And Signal Alignment, Inc.  Binder identification 
US7817745B2 (en) *  20050602  20101019  Adaptive Spectrum And Signal Alignment, Inc.  Tonal precoding 

2009
 20090824 KR KR1020117003726A patent/KR101600332B1/en not_active IP Right Cessation
 20090824 EP EP09810359.1A patent/EP2319191A4/en not_active Withdrawn
 20090824 US US12/583,694 patent/US8427933B2/en active Active
 20090824 WO PCT/US2009/004838 patent/WO2010024884A1/en active Application Filing
Patent Citations (9)
Publication number  Priority date  Publication date  Assignee  Title 

US20020042836A1 (en) *  20000407  20020411  Mallory Tracy D.  Method of enhancing network transmission on a priorityenabled framebased communications network 
US20030206559A1 (en) *  20000407  20031106  Trachewsky Jason Alexander  Method of determining a start of a transmitted frame in a framebased communications network 
US20020027985A1 (en) *  20000612  20020307  Farrokh RashidFarrokhi  Parallel processing for multipleinput, multipleoutput, DSL systems 
US20080198909A1 (en) *  20030908  20080821  Michail Konstantinos Tsatsanis  Efficient multiple input multiple output signal processing method and apparatus 
US7400710B2 (en) *  20040226  20080715  Conexant, Inc.  MIMO dynamic PSD allocation for DSL networks 
US20060259535A1 (en) *  20050510  20061116  Adaptive Spectrum And Signal Alignment, Inc.  Tonal rotors 
US20070230548A1 (en) *  20060317  20071004  Broadcom Corporation  Digital subscriber line noise mitigation techniques, and applications thereof 
US20090059780A1 (en) *  20070831  20090305  De Lind Van Wijngaarden Adriaan J  Method and apparatus for selftuning precoder 
US20090116582A1 (en) *  20071102  20090507  Alexei Ashikhmin  Interpolation method and apparatus for increasing efficiency of crosstalk estimation 
Cited By (28)
Publication number  Priority date  Publication date  Assignee  Title 

US20100034374A1 (en) *  20061020  20100211  Oehman Hans  Vectored digital subscriber line system having modular vectoring arrangements 
US8126137B2 (en) *  20061020  20120228  Ericsson Ab  Vectored digital subscriber line system having modular vectoring arrangements 
US8923454B2 (en) *  20090827  20141230  Huawei Technologies Co., Ltd.  Method, apparatus and system for eliminating aliasing noise in multicarrier modulation system 
US20120114063A1 (en) *  20090827  20120510  Huawei Technologies Co., Ltd.  Method, apparatus and system for eliminating aliasing noise in multicarrier modulation system 
US20110235759A1 (en) *  20100323  20110929  Ikanos Communications, Inc.  Systems and Methods for Implementing a MultiSensor Receiver in a DSM3 Environment 
WO2011119519A1 (en)  20100323  20110929  Ikanos Technology Ltd.  Systems and methods for implementing a multisensor receiver in a dsm3 environment 
CN102884729A (en) *  20100323  20130116  伊卡诺斯通讯公司  Systems and methods for implementing a multisensor receiver in a DSM3 environment 
US8576690B2 (en)  20100323  20131105  Ikanos Communications, Inc.  Systems and methods for implementing a multisensor receiver in a DSM3 environment 
TWI483556B (en) *  20100323  20150501  Ikanos Communications Inc  Systems and methods for implementing a multisensor receiver in a dsm3 environment 
US20140056396A1 (en) *  20100323  20140227  Ikanos Communications, Inc.  Systems and methods for implementing a multisensor receiver in a dsm3 environment 
US9184774B2 (en) *  20100323  20151110  Ikanos Communications, Inc.  Systems and methods for implementing a multisensor receiver in a DSM3 environment 
US20120026926A1 (en) *  20100531  20120202  Rudi Frenzel  Low power mode for vectored data transmission 
US9749008B2 (en) *  20100531  20170829  Lantiq BeteiligungsGmbH & Co. KG  Low power mode for vectored data transmission 
US20140314071A1 (en) *  20111231  20141023  Huawei Technologies Co., Ltd.  Method, Apparatus and System for MultiCarrier OFDM Duplex Transmission 
US10382189B2 (en) *  20111231  20190813  Huawei Technologies Co., Ltd.  Method, apparatus and system for multicarrier OFDM duplex transmission 
US9686035B2 (en) *  20120723  20170620  Lantiq Deutschland Gmbh  Spectrum management and timing optimization over multiple distribution points 
US20140056312A1 (en) *  20120723  20140227  Lantiq Deutschland Gmbh  Spectrum Management and Timing Optimization Over Multiple Distribution Points 
US10033430B2 (en)  20121030  20180724  Lantiq Deutschland Gmbh  Spectrum management 
US10116350B2 (en)  20121030  20181030  Lantiq Deutschland Gmbh  Channel estimation 
US9584181B2 (en) *  20121030  20170228  Lantiq Deutschland Gmbh  Channel estimation 
US20140119419A1 (en) *  20121030  20140501  Lantiq Deutschland Gmbh  Channel Estimation 
US9756178B2 (en)  20140724  20170905  Huawei Technologies Co., Ltd.  Crosstalk estimation method, apparatus, and system 
CN105453501A (en) *  20140724  20160330  华为技术有限公司  Crosstalk estimation method, device, and system 
EP3154230A4 (en) *  20140724  20170628  Huawei Technologies Co. Ltd.  Crosstalk estimation method, device, and system 
US20160248475A1 (en) *  20150224  20160825  Lantiq BeteiligungsGmbH & Co. KG  Crosstalk mitigation 
US10141976B2 (en) *  20150224  20181127  Lantiq BeteiligungsGmbH & Co. KG  Crosstalk mitigation 
EP3255807A4 (en) *  20150429  20171220  Huawei Technologies Co., Ltd.  Method, apparatus and system for sending signal 
CN109728837A (en) *  20171030  20190507  华为技术有限公司  A kind of method, apparatus and system for offsetting crosstalk signal 
Also Published As
Publication number  Publication date 

EP2319191A4 (en)  20170628 
KR20110050641A (en)  20110516 
WO2010024884A1 (en)  20100304 
EP2319191A1 (en)  20110511 
US8427933B2 (en)  20130423 
KR101600332B1 (en)  20160307 
Similar Documents
Publication  Publication Date  Title 

EP1733550B1 (en)  High speed multiple loop dsl system  
US7394752B2 (en)  Joint reduction of NEXT and FEXT in xDSL systems  
US6999504B1 (en)  System and method for canceling crosstalk  
CN101005323B (en)  Dynamic digital communication system control  
US7561627B2 (en)  Method and system for channel equalization and crosstalk estimation in a multicarrier data transmission system  
JP3679722B2 (en)  Enhanced bit loading for multicarrier communication channels  
EP2055014B1 (en)  Systems and methods for mimo precoding in an xdsl system  
JP3692044B2 (en)  Equalizer training for ADSL transceivers in TCMISDN crosstalk environment  
US8139658B2 (en)  Method and system for providing a time equalizer for multiline transmission in communication systems  
Cioffi et al.  Veryhighspeed digital subscriber lines  
US5887032A (en)  Method and apparatus for crosstalk cancellation  
US6266367B1 (en)  Combined echo canceller and time domain equalizer  
EP1109328B1 (en)  DSL transmission system with farend crosstalk compensation  
US8179774B2 (en)  Crosstalk coefficient updating in vector transmission  
US6404806B1 (en)  Method and apparatus for timedomain equalization in FDMbased discrete multitone modems  
US8537912B2 (en)  Extremely high speed broadband access over copper pairs  
KR101567120B1 (en)  Alien interference removal in vectored dsl  
JP3469558B2 (en)  Method and apparatus for timing recovery in an ADSL receiver under TCMISDN crosstalk  
US6324212B1 (en)  Apparatus using low spectrum selectively for providing both ADSL and POTS service  
EP1476951B1 (en)  Crosstalk mitigation in a modem pool environment  
US20050276340A1 (en)  Method and system for determining symbol boundary timing in a multicarrier data transmission system  
US6965657B1 (en)  Method and apparatus for interference cancellation in shared communication mediums  
JPH09289476A (en)  Interference reduction in communication system  
US8270524B2 (en)  Method and apparatus for interference postcompensation using a bandwidthadaptive postcoder interface  
JP5808507B2 (en)  Method and system for reducing crosstalk 
Legal Events
Date  Code  Title  Description 

AS  Assignment 
Owner name: IKANOS COMMUNICATIONS, INC.,CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:SCHELSTRAETE, SIGURD;HEIDARI, SAM;REEL/FRAME:023528/0206 Effective date: 20091110 Owner name: IKANOS COMMUNICATIONS, INC., CALIFORNIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:SCHELSTRAETE, SIGURD;HEIDARI, SAM;REEL/FRAME:023528/0206 Effective date: 20091110 

STCF  Information on status: patent grant 
Free format text: PATENTED CASE 

CC  Certificate of correction  
AS  Assignment 
Owner name: ALCATELLUCENT USA, INC., NEW JERSEY Free format text: NOTICE OF GRANT OF SECURITY INTEREST IN PATENTS;ASSIGNOR:IKANOS COMMUNICATIONS, INC.;REEL/FRAME:035581/0710 Effective date: 20150430 

AS  Assignment 
Owner name: SILICON VALLEY BANK, CALIFORNIA Free format text: SECURITY INTEREST;ASSIGNOR:IKANOS COMMUNICATIONS, INC.;REEL/FRAME:035874/0351 Effective date: 20150602 

AS  Assignment 
Owner name: IKANOS COMMUNICATIONS, INC., CALIFORNIA Free format text: RELEASE BY SECURED PARTY;ASSIGNOR:ALCATELLUCENT USA, INC.;REEL/FRAME:036732/0876 Effective date: 20150929 Owner name: IKANOS COMMUNICATIONS, INC., CALIFORNIA Free format text: RELEASE BY SECURED PARTY;ASSIGNOR:SILICON VALLEY BANK;REEL/FRAME:036733/0031 Effective date: 20150930 

FPAY  Fee payment 
Year of fee payment: 4 